CGD Policy Paper 019April 2013

Impact of Conditional Cash Transfers on Maternal and Newborn Health

Conditional cash transfers (CCTs) have been shown to increase health service utilization among the poorest, but little is written on the effects of such programs on maternal and newborn health.

We carry out a systematic review of studies on CCTs that report maternal and newborn health outcomes, including studies from eight countries. We find that CCTs have increased antenatal visits, skilled attendance at birth, delivery at a health facility, and tetanus toxoid vaccination for mothers, and reduced the incidence of low birth weight. The programs have not had a significant

impact on fertility or Caesarean sections while impact on maternal and newborn mortality has not been well documented thus far.

Given these positive effects, we discuss priorities for future investment in CCT programs for maternal and newborn health, noting gaps in knowledge and providing recommendations for better design and evaluation of such programs. We recommend more rigorous impact evaluations that document impact pathways, look into outcomes, and take factors such as cost-effectiveness into account.

Amanda Glassman, Denizhan Duran, and Marge Koblinsky

Center for Global Development1800 Massachusetts Ave NWThird FloorWashington DC 20036202-416-4000 www.cgdev.org

This work is made available under the terms of the Creative Commons Attribution-NonCommercial 3.0 license.

Abstract

Amanda Glassman, Denizhan Duran, and Marge Koblinsky. 2013. “Impact of Conditional Cash Transfers on Maternal and Newborn Health.” CGD Policy Paper 019. Washington DC: Center for Global Development. www.cgdev.org/publication/impact-conditional-cash

CGD is grateful for contributions from the the William and Flora Hewlett Foundation, the UK Department for International Development, and the Swedish Ministry of Foreign Affairs in support of this work.

Contents

1. Introduction ............................................................................................................................. 1

2. Conditional Cash Transfers and Maternal Health ............................................................. 2

3. Methods .................................................................................................................................... 4

4. Results ....................................................................................................................................... 5

Perinatal, neonatal and maternal mortality .............................................................................. 5

Low birth weight .......................................................................................................................... 5

Adequate antenatal monitoring ................................................................................................. 5

Births attended by skilled personnel ......................................................................................... 8

Births in health facilities .............................................................................................................. 8

Cesarean section ........................................................................................................................... 8

Tetanus toxoid vaccination for mothers .................................................................................. 8

Post-partum visits ........................................................................................................................ 8

Fertility ........................................................................................................................................... 9

Other outcomes: Contraceptive use, HIV status .................................................................... 9

5. Discussion ................................................................................................................................ 9

6. Conclusions and Recommendations .................................................................................. 13

Annex 1. Features of CCT programs in countries where maternal health outcomes were

measured .......................................................................................................................................... 15

Annex 2. Forest plot tables and graphs for interventions ....................................................... 16

Annex 3. Overall table of reviewed effects ................................................................................ 21

Annex 4. Definition of terms ....................................................................................................... 23

References ....................................................................................................................................... 24

Amanda Glassman: Director of Global Health Policy and Senior Fellow, Center for Global

Development. Denizhan Duran: Research Assistant, Center for Global Development; Marge

Koblinsky: Senior Maternal Health Advisor, USAID. Corresponding author:

aglassman@cgdev.org.

Acknowledgments

This paper was written with contributions from the members of the US Government

Evidence Review Team 2, which worked on this paper during an April 2012 meeting. The

members include the authors of this policy paper, as well as Daniel Singer (NICHD), Rachel

Sturke (NIH), Gustavo Angeles (University of North Carolina Chapel Hill), Jodi Charles

(USAID), Bob Emrey (USAID), Lisa Fleisher (USAID), Joanne Gleason (Population

Council), Howard Goldberg (CDC), Winnie Mwebsa (Save the Children), Kelly Saldana

(USAID), Kristina Yarrow (USAID). The authors thank Justin Sandefur and an external

reviewer for helpful comments.

1

1. Introduction

Given the slow decline in maternal and newborn mortality since 1990, the achievement of

Millennium Development Goals 4 and 5 by 2015 is unlikely. Most of these deaths occur in

the intra-partum and immediate post-partum period largely from preventable causes [1] [2].

Annually about 60 million women give birth outside of health facilities, mainly at home, and

52 million without a skilled birth attendant [3]. Family planning needs are satisfied for only

about 50% of women [4] and fertility rates are still as high as 4 live births per woman in low-

income countries [5]. Further, skilled birth attendance, use of antenatal care and satisfaction

with family planning are the most inequitably distributed of 12 key maternal, newborn and

child health interventions studied in low and middle income countries (LMIC) with poorer

women facing higher barriers to access [4].

The reasons behind the limited use of health services by the poor are myriad, and occur on

both the supply (providers) and demand (households, women) sides. On the demand side,

poverty, poor health status, illiteracy, language, customs, lack of information regarding the

availability of health services and providers, and limited control over household resources

and decision-making all play a role in limiting access to care. On the supply side, poor quality

provision (both antenatal and obstetric care), mistreatment or sociocultural insensitivity,

absence of a trained attendant at delivery, inadequate referral systems for emergency

obstetric care, inadequate or lack of transportation facilities, and absence of or poor linkages

of health centers with communities are barriers to utilization [6].

Increasingly, maternal and newborn health (MNH) experts are exploring ways in which

demand-side barriers – the barriers women and their families face to seeking care – can be

overcome. Over the past decades, community-based programs to increase utilization of

MNH care have been piloted, most notably through community mobilization, behavior

change communications, and volunteer outreach efforts, among others. These programs

have in general been small-scale and have not entirely solved the major barrier to care

seeking: the direct and indirect financial costs associated with seeking care [7].

Conditional cash transfers (CCT) are one type of demand-side program that has been used

to overcome cost barriers. CCT are social programs that condition regular cash payments to

poor households on use of certain health services and school attendance. CCT programs

have two main objectives: first, to provide a safety net to increase and smooth the

consumption of the extreme poor (alleviating short-term poverty), and second, to increase

the human capital investment of poor households (alleviating long-term poverty). Payments

are usually provided to women and compliance with conditions is verified by the program,

and transfer sizes are generally intended to close the gap between average consumption in

the bottom quintile of the income distribution and the extreme poverty line. Initially based

in Latin America, CCT programs now operate around the world, and are regarded as

successful social protection strategies.

2

Although few CCT programs have explicitly targeted the improvement of maternal and

newborn health, many of the “broad” programs included conditionality, associated supply-

side strengthening and/or educational talks related to MNH, and many impact evaluations

have measured the effects of CCT on MNH interventions and outcomes. Further, as

programs that relax a household’s budget constraint, CCT can be expected to affect

household spending choices in general, including the potential to improve MNH. Other

systematic reviews have documented the effects of CCT on child health care utilization and

nutritional status [8], yet no paper has directly reviewed the evidence on CCT impact on

MNH or use of appropriate MNH services.

This paper will fill this gap, setting out the hypothesized channels through which CCT

impact on MNH may occur, synthesizing the empirical evidence on the impact of CCT on

MNH interventions and outcomes, discussing issues emerging from the evidence synthesis,

and providing recommendations for the future. Specifically the questions we address are:

Are CCT linked positively or negatively to maternal and neonatal health outcomes,

use or provision of maternal health services, or to care-seeking behavior by women?

What are the contextual factors that mediate the effectiveness of CCT?

2. Conditional Cash Transfers and Maternal Health

Most CCT programs are broad, aiming to alleviate poverty and increase human capital

through transfers that are conditioned on a combination of school attendance, use of well-

child visits, vaccination, and/or use of nutritional supplements. Examples of “broad” CCT

include Mexico’s Oportunidades, Colombia’s Familias en Accion, Nigaragua’s Red de Proteccion

Social, Honduras’ PRAF, Turkey’s SRMP, among others. However, “narrow” CCT –

programs that transfer cash only for the utilization of specific services- are becoming more

common; for example, India’s Janani Suraksha Yojana (JSY) and Nepal’s Safe Delivery

Incentive Program (SDIP) target MNH improvements specifically. However, unlike the

broad CCT, the “narrow” programs like JSY and SDIP do not always or only target low-

income groups. In JSY, a mix of geographical and income targeting is used to induce

pregnant women to seek care, while in Nepal, cash incentives are offered to all pregnant

women [9].

Although programs differ in their specific design features, CCT usually share the following

key features (see figure 1):

Cash transfers that are conditioned on the utilization of a service, as mandated

under the program ie health, education, nutrition

Health information, education and communication (IEC) efforts

Ex-ante identification (“targeting”) of recipient communities or households, using a

variety of criteria

Verification of compliance with conditions

3

Figure 1. Conceptual Framework of CCT Programs

These features of CCT programs are potentially associated with MNH outputs and

outcomes on both the demand- and supply- sides. On the demand-side, household-level

outputs include improved nutrition and feeding, as well as better newborn care, such as

exclusive breastfeeding, delayed bathing, warmth and cord care. At the health system level,

demand-side outputs involve care-seeking behaviors, such as use of antenatal care (ANC),

use of facilities for birth, and use of a skilled birth attendant. Increased demand for services

may also trigger improvements in the supply of services via greater and improved provider

responsiveness (e.g. less absenteeism).

To support such improvements, some conditional cash transfer programs include

components that support the supply-side, such as strengthening health services in program

areas. In its initial phase, Nicaragua’s RPS, for example, contracted non-governmental

organizations to provide an essential package of services to CCT beneficiaries and non-

beneficiaries in intervention communities [10]. Similarly, India’s JSY program has a supply-

side component, including incentive payments to community level health workers for

bringing pregnant women to a designated facility for delivery [11].

Together demand- and supply-side outputs –mediated by contextual factors- are expected to

jointly generate improved newborn outcomes such higher birth weights and survival (both

perinatal and neonatal), while maternal health outcomes could include anemia, fertility,

survival and complications during pregnancy and birth.

4

3. Methods

Our initial search of “conditional cash transfers & maternal health” resulted in 5,800 results

on Google Scholar, of which 470 documents remained after duplicates were removed. A call

for relevant papers led to 26 additional documents. A total of 65 documents remained after

the screen was applied (see Annex A), categorized according to financial incentive and

outcome/s reported. Of these, 9 articles applied to CCT.

In addition, a search for non-financial strategies to increase care seeking for MNH services

was conducted using the search terms, “care seeking and maternal”, “care seeking and

newborn”, “care seeking and postpartum family planning” and search engines (Medline,

Cochrane Collection); 72 hits were received and abstracts reviewed. Final articles selected

numbered 24 and included primarily Cochrane reviews, systematic reviews, with single

published papers that explored specific issues of interest (e.g., birth preparedness

complication readiness).

The characteristics of the CCT programs that have MNH outcomes are reported in Annex 1.

Although we do not have more than one study for each program, the studies included are

well-designed impact evaluations with experimental or quasi-experimental designs, with

output measures that are relatively comparable and consistent across different studies.

Of the CCT studies that report rigorously calculated impacts, outcome variables that were

common across at least two studies were identified and baseline values, effect sizes (reported

as the average treatment effect, or the difference between treatment and control groups),

standard errors, significance, sample sizes and scope of the program, defined as the ratio of

beneficiaries over the total population, are reported.

Forest plots are used to depict effect sizes and pooled estimates. In order to mitigate non-

comparability, we use a DerSimonian – Laird random-effects model, a widely used method

to construct forest plots. Assuming heterogeneity between studies, this method uses a non-

iterative method to estimate the inter-study treatment effect variance, without making any

assumptions regarding the distribution of within-study or between-study effects. These

estimations are generated with the metaan function on Stata 12. The final forest plots show

individual effect sizes with confidence intervals, as well as a final average effect size; the size

of the boxes shows the significance of the effect, where larger boxes indicate a wider range

and larger confidence intervals. We report pooled average effects only when two or more

studies are available.In annex 2, we show forest plots for each outcome variable across

available studies. In annex 3, we also provide a table of effect size and confidence intervals

for all MNH-relevant variables reported in the studies.

5

4. Results

In this section, we describe the results of qualifying CCT studies on various MNH indicators

summarized in Table 1, next page (see annex 4 for how the studies define these terms).

These same results are depicted graphically in forest plots in Annex 2. Results on MNH

mortality are also discussed. We aggregate results from CCT programs in 8 countries with 14

evaluations.

Perinatal, neonatal and maternal mortality

Studies do not report comparable measures of mortality and were thus omitted from the

summary table. However, three studies look at mortality measures. The official evaluation of

Oportunidades reports an 11% decline in maternal mortality in regions that had at least one

locality incorporated in the Oportunidades program (RR 0.89, IC 0.82, 0.95) [23] . Lim et al

report large declines in perinatal and neonatal deaths associated with India’s JSY, although

findings for maternal death were insignificant [14]. Powell-Jackson et al (2010) report a very

small, insignificant decline in neonatal mortality in Nepal [15]. However, estimating impact

on maternal mortality is prone to measurement errors and underestimation given issues in

data collection, so we do not aggregate these results (see box 1 on criticisms regarding the

JSY impact evaluation).

Low birth weight

Two studies analyze the effect of CCT on the incidence of low birth weight, in Mexico and

Uruguay [16], [18]. Both studies report a small but significant decline in the incidence of low

birth weight: in Mexico, the proportion of infants born with low birth weight declined by

4.6%, and in Uruguay, by 1.5%.1 An unpublished job market paper from Indonesia found

that the CCT program did not have an impact on low birth weight or other birth outcomes

[27].

Adequate antenatal monitoring

Most studies report on the effect of the CCT program on the average number of antenatal

visits. Four studies report positive and significant increases in the average number of

beneficiaries that received at least 5 antenatal visits among beneficiaries compared to non-

beneficiaries, ranging from an 8 percentage point difference in Mexico to a 19 percentage

point increase in Honduras. The Honduran program is also the only program that included a

specific conditionality related to antenatal care use; other programs (El Salvador, Mexico and

Guatemala) only required preventive health care utilization by children while the remaining

programs only conditioned facility births [13]. Two programs - El Salvador and Nepal -

1 Overall effects here refer to the effect sizes that are calculated in the DerSimonian – Laird random effects

models (see annex 2 for weights assigned by this method).

Table 1. Maternal Health CCT Programs, Effect Sizes

Study & Country Baseline Effect Size Standard Error

Significance Sample Size % of Population as Beneficiary

Adequate prenatal monitoring2 de Brauw and Peterman (2011) (El Salvador) [12] 0.768 -0.065 0.072 NS 494 0.131 Morris, Flores, Olinto and Medina (2004) (Honduras) [13] 0.379 0.187 0.060 *** 313 0.150

Lim et al (2010) (India) [14] 0.536 0.107 0.008 *** 182869 0.100

Powell-Jackson et al (2010) (Nepal) [15] 1.235 -0.046 0.061 NS 5901 No targeting Barber and Gertler (2009) (Mexico) [16] 0.612 0.081 0.026 *** 892 0.180 IDB; Gutierrez et al (2011) (Guatemala) [17] 2.69 0.11 0.067 ** 1163 0.057 Amarante et al (2011) (Uruguay) [18] 6.53 0.144 0.059 ** 67863 0.100

Births attended by skilled personnel de Brauw and Peterman (2011) (El Salvador) [12] 0.738 0.123 0.070 * 536 0.131 Lim et al (2010) (India) [14] 0.593 0.366 0.006 *** 182869 0.100 Powell-Jackson et al (2010) (Nepal) [15] 0.225 0.052 0.016 *** 5901 No targeting Urquieta et al (2009) (Mexico) [19] 0.305 0.114 0.048 ** 860 0.180 IDB; Gutierrez et al (2011) (Guatemala) [17] 0.105 0.04 0.031 * 1006 0.057 Amarante et al (2011) (Uruguay) [18] 0.49 -0.002 0.009 NS 68855

Tetanus toxoid for mother Morris, Flores, Olinto and Medina (2004) (Honduras) [13] 0.563 0.042 0.071 NS 313 0.150 Barber and Gertler (2009) (Mexico) [16] 0.924 0.368 0.300 NS 892 0.180

Gave birth in hospital Powell-Jackson et al (2010) (Nepal) [15] 0.106 0.04 0.015 *** 5901 No targeting Lim et al (2010) (India) [14] 0.541 0.435 0.006 *** 182869 0.100 de Brauw and Peterman (2011) (El Salvador) [12] 0.733 0.153 0.076 * 530 0.131

Post-partum checkups/visits after birth Morris, Flores, Olinto and Medina (2004) (Honduras)3 [13] 0.178 -0.056 0.052 NS 311 0.150 de Brauw and Peterman (2011) (El Salvador) [12] 0.259 -0.059 0.100 NS 478 0.131

Contraceptives Feldman et al (2009) (Mexico) [20] 0.37 0.16 0.097 ** 16462 0.180 Lamadrid-Figueroa et al (2010) (Mexico) [21] 0.39 0.049 0.036 NS 2239 0.180 Powell-Jackson et al (2010) (Nepal) [15] 0.025 0.012 0.006 * 5901 No targeting Barber and Gertler (2009) (Mexico) [16] 0.145 0.051 0.031 ** 979 0.180

Fertility Stecklov et al (2007) (Honduras) [22] 0.172 0.039 0.014 ** 12677 0.150 Stecklov et al (2007) (Nicaragua) [22] 0.108 -0.011 0.067 NS 4885 0.027 Stecklov et al (2007) (Mexico) [22] 0.179 -0.003 0.003 NS 17634 0.180 Amarante et al (2011) (Uruguay) [18] N/A 0.001 0.000 *** 1037793 0.100

Low birthweight Barber and Gertler (2009) (Mexico) [16] N/A -0.046 0.096 ** 804 0.180 Amarante et al (2011) (Uruguay) [18] 0.102 -0.015 0.005 *** 68858 0.100

1. Defined as 5 or more visits in every paper except for India, where it’s defined as 3 or more visits. Numbers in italics are number of visits; other numbers are percentage of the population who have received adequate prenatal monitoring. 2. Defined as a 10-day post-partum checkup Legend: *: Significant in the 10% level, **: Significant in the 5% level, ***: Significant in the 1% level, NS: Not significant

7

BOX 1. INDIA’S JANANI SURAKSHA YOJANA (JSY) PROGRAM: ISSUES WITH IMPACT

EVALUATIONS

India’s JSY is the largest CCT program in the world and specifically targets MNH. This,

coupled with the fact that India has the highest number of maternal deaths in the world,

makes JSY’s evaluation extremely important. Lim et al (2010) [14], the only published

impact evaluation of the JSY program to date (between 2002-2004 and 2007-2008),

reports positive results for service uptake as well as neonatal mortality.

However, new studies challenge some of these findings. An unpublished evaluation

by Mazumdar, Mills and Powell-Jackson (2012) [24] finds similar results for JSY increases

on facility deliveries, also reporting that the program was more effective for less

educated, poor and ethnically marginalized women. The study also finds increases in

breastfeeding and less use of private health providers. Going beyond the positive impact

on service uptake, however, the Mazumdar et al study finds that JSY increases fertility

and does not have an effect on antenatal care or neonatal mortality.

Other critics of the program point to gaps in the evaluation. A letter published in the

Lancet by Das et al (2011) points to problems in the enforcement of conditions,

inconsistencies in the implementation of the program between states, and problems in

recording program enrollment status due to an ambiguous question in the household

survey [25]. Beyond these concerns, a process evaluation of the program by Devadasan

et al (2008) [26] raised issues of women receiving only a portion of the promised cash

transfers, as well as the transfer going to women who delivered at home.

These concerns about the world’s largest CCT program show the need for a

systematized, more rigorous design for impact evaluations, as well as more attention to

process evaluation.

reported a small decline in the average number of antenatal visits, but these results were

insignificant [12, 15]. The Indian JSY program conditioned three or more visits, and reports

an 11 percentage point increase. It is important to note the relatively short timeframe of

these programs; the Indian JSY program generated such an outcome in a span of two years

[14]. Overall, the effect is an increase in the uptake of adequate prenatal monitoring services

by 8.4%.

A caveat of these studies is their focus on quantity, not quality, of antenatal care, particularly

important in light of the null impact that antenatal care utilization on its own can have on

maternal and birth outcomes. In the only study that examines quality, Barber and Gertler

(2009) report positive effects of Mexico's Oportunidades program on the number of Ministry

of Health-recommended prenatal procedures provided during antenatal visits, as well as the

number of iron supplements provided [16].

8

Births attended by skilled personnel

Six studies reported on the effect of the CCT program on whether a woman's last birth was

attended by skilled personnel. For this variable, every study reported positive and significant

outcomes, from a low of a 4 percentage point difference between beneficiaries and non-

beneficiaries in Guatemala to a high of 37 percentage point difference in India [14]. The

overall effect is a 12% increase; however, this impact has a wide confidence interval given

the high variance between effect sizes. A forthcoming study from Indonesia reports a 45%

increase in skilled delivery, but is not included here because it does not report a standard

error for its effect size [27].

Births in health facilities

Three studies reported on the effect of the CCT program on whether a woman's last birth

occurred in a hospital. Each study reported positive and significant effects, and effect sizes

much larger than those reported for other outcomes. In Nepal, there was a 4 percentage

point difference between beneficiaries and non-beneficiaries [15], while Lim et al reported a

43.5 percentage point difference in India [14]. Due to this high variance, the overall effect of

21% has a very large confidence interval.

Cesarean section

Two studies reported on the CCT effect on cesarean section at last birth among beneficiaries

and non-beneficiaries. Both studies reported positive and significant effects; in Nepal, there

was a one percentage point difference between the intervention and control groups [15],

while this difference was 5 percentage points in Mexico [16]– the overall effect is a 2%

increase. However, the Mexico baseline rate among beneficiaries and their controls was

already 15% of all births, the top of the WHO-recommended level, so it is not possible to

interpret whether this increase is consistent with MNH recommendations for better

outcomes.

Tetanus toxoid vaccination for mothers

Two studies reported on the impact of the CCT on the probability that a mother would

receive a tetanus toxoid vaccination, an intervention that is essential to ensure survival of

both mother and baby in LMIC, especially where there is a large share of home births [13,

28]. While the effects were positive, neither result was statistically significant. Given this, the

overall effect is an 8% increase (average is between a 37% increase in Mexico and 4%

increase in Honduras).

Post-partum visits

Two studies measured the effect of the CCT program on post-partum visits, considered

critical for both mother and newborn, especially in the immediate 48 hours through the first

9

week following birth [13], [12]. Both studies found negative but insignificant results on this

variable, with an overall effect size of 6% decline in post-partum visits.

Fertility

A study looking at the changes in fertility rates from Honduras, Nicaragua, Mexico and

Uruguay reports impact on age-specific and total fertility rates. The overall effect is

negligible, with a .2% increase, and range from a 4% increase in Honduras to a 1% decrease

in Nicaragua [22]. Honduran women were provided per-child benefits from birth, and this

design may have resulted in a change in the tempo of fertility or perhaps total fertility among

beneficiaries. A study on JSY finds that fertility increased by 1.1 percentage points, but

reports non-comparable outcomes and is therefore not included in table 2 [24].2 A CCT

evaluation in Pakistan finds that a beneficiary's probability of giving birth was 8 percentage

points less than a non-beneficiary; the beneficiaries were more likely to have a smaller

number of children and more likely to be older at marriage [29]. In a Malawi CCT (where

conditionality is related to school attendance, not health care use), adolescent beneficiaries

were significantly less likely to become pregnant [30].

Other outcomes: Contraceptive use, HIV status

Only one program -Mexico's Oportunidades- reported on contraceptive use, finding that

beneficiaries were 16 percentage points more likely to use a modern contraceptive method

than non-beneficiaries . Another analysis of Mexico’s Oportunidades looks at heterogeneous

effects, finding a small and insignificant effect on contraceptive use [21]. There are also

results on risky sexual behavior from both Oportunidades and the Malawi CCT: the Malawi

study found that those in the CCT program were 0.29 times less likely to have HIV [31].

5. Discussion

The review of effect sizes suggests that CCT can reduce barriers to MNH service utilization

such as prenatal monitoring, skilled attendance at birth and use of facility for birth. Further,

CCT may have an impact on the incidence of low birth weight, as well as the more distal

outcomes of fertility and mortality. However, more studies are required to make a definitive

statement.

In this section, we discuss the potential impact pathways for CCT programs and discuss how

CCT programs compare to non-financial interventions in terms of outcomes and cost

effectiveness.

2 The Mazumdar et al study [24] measures fertility by assuming that respondents are pregnant for 6 months

when they respond to the questionnaire, whereas the Stecklov study [22] reports fertility based on baseline

surveys that were conducted before the program & estimates fertility using differences-in-differences.

10

Interpreting these results is contingent upon understanding the channels leading to results.

Studies hypothesize that the impact of CCT on maternal and newborn health is channeled

through one or more of the following channels:

(i) Income effect: Household income increases and/or share of income that

women control increases, thus there is more disposable income to spend on

health;

(ii) Conditioning cash specifically on the usage of maternal health services

(iii) Removal of costs associated with service utilization via subsidies for defined

health benefit plans that include maternal care

(iv) Knowledge effects resulting from health or nutrition training/talks, especially in

Latin American programs.

It is difficult, however, to define specific causal pathways and link CCT design features

directly to impact, given that the evaluations were not designed to measure the effects of

these channels. It is especially difficult to track this as most CCT programs reviewed here

have a broad scope and do not focus on maternal and newborn health.

Another key issue that would determine impact channels is whether utilization is related to

actual health outcomes, such as low birth weight and decreased maternal and neonatal

mortality. Although the studies we report from India, Mexico and Uruguay point out to

better outcomes, more needs to be done in the field of connecting utilization to outcome,

and future studies might show that encouraging utilization might not necessarily lead to

better health outcomes.

While evidence suggests that CCT programs can improve utilization of maternal services and

may improve maternal and newborn outcomes, successful implementation depends on a

range of contextual factors. Their success is dependent on their ability to counter barriers

to accessing care, and preferably improve quality care. When the barrier is poverty, CCT

provide a direct incentive—but rarely is poverty the only barrier. Care seeking and health

outcomes are determined by the interplay of social, cultural, health system factors as well as

those that are economic. As these factors vary by context, a CCT that is successful in one

context may be unsuccessful in another, with the difference attributable to factors that are

not typically assessed during a program evaluation, such as cultural factors or supply-side

constraints.

Some contextual enabling factors underpinning the effectiveness of CCTs include the

following:

Macroeconomic stability and economic growth where financial strength allows for

increased investments in social infrastructure, especially in education and health.

Enhanced infrastructure which generates enough supply to meet the additional

demand. This includes skilled providers, adequate processes and systems, systematic

11

reporting, monitoring and evaluation, training and supervision of health staff and

appropriate drug supplies and transport.

Strong information systems including adequate household data to enable effective

identification of beneficiaries.

Strong targeting mechanisms to ensure high coverage and minimize the number of

non-poor who benefit from the program.

Conversely many contextual factors can be persistent barriers to the implementation of

successful CCT programs:

Poor Infrastructure: CCT can be used to improve access to transportation, but will be

ineffective when the transportation infrastructure is so poor that cost is not the

major determinant (e.g. areas where there are insufficient vehicles, fuel shortages,

roads or where conditions such as floods make transportation impossible).

Quality of Care: CCT can incentivize women to seek antenatal, delivery and

postpartum care, but improvement of maternal and newborn health outcomes is

dependent on the quality of care delivered. Quality of care is often impaired by drug

stock-outs, use of counterfeit drugs, poor training or retention/absenteeism of

health care staff, lack of necessary equipment, etc. Low funding, mismanagement of

funds, lack of leadership and management, has left many hospitals short of staff and

vital supplies essential for delivery of quality care to women.

Political Considerations: Slater et al. (2008) argue that the political context for cash

transfers is potentially complex, but understanding it is an essential step in making

the context more favorable for transfers [32]. The overall political feasibility of cash

transfers depends on, for example, the type and extent of political commitment to

poverty reduction and the overall availability of resources for social transfers.

Political acceptability on a more day-to-day basis will depend on the size and cost of

administrative effort to implement cash transfers, but also on the perceptions of the

electorate: prejudices against perceived handouts may limit cash transfers to “cash

for work” modalities, in which the poor are seen to be earning transfers.

Societal and gender norms: In some areas strong cultural norms impede women from

seeking quality maternal health services. These include preference for traditional

healers and prohibitions on a woman traveling far from her home without

accompaniment by an appropriate family member. In these instances, a CCT

program is only effective when its perceived economic and health benefits outweigh

the perceived cost of crossing traditional and societal norms. Strong gender

stratification can influence the success of CCT. While the activities of poverty

alleviation programs can contribute to women’s empowerment and provide women

with greater autonomy, bargaining power and a larger social network, backlash may

be possible, including conflict and even violence as husbands or others preferring

traditional gender norms protest the women’s greater power. However, in

quantitative and qualitative evaluations of CCT to date, there has been no evidence

of these negative effects to date [33].

12

Another important dimension of CCT is the sustainability of the behavior change desired

and whether and how long improved behavior relies on the existence of the financial

incentive. As CCT are relatively new instruments for change, such behavior change regarding

the use of maternal health services has not been examined. Some evidence suggests that

there can be a learning effect whereby women with greater exposure time to a CCT program

engage in greater utilization of maternal health services. For example, an increase in the last

delivery attended by a physician/nurse vs. a traditional midwife in Mexico was reported

although the CCT only specifies use of adequate antenatal care, not facility delivery or use of

a skilled birth attendant [34] These studies do not, however, examine the absence of the

financial incentive on behavior. More robust research in this area is needed to provide the

evidence on sustainability of behavior changes.

The sustainability of behavior change also depends on the financial sustainability for

improved MNH outcomes depend on the financial sustainability of programs, at least

initially. In low-income countries, CCT programs are often funded by external donors,

whereas in middle-income countries, national governments are more likely to fund the

program directly. In LIC country settings, donor sustainability is a persistent issue, and in

both LIC and MIC settings, political support plays a major role in sustainability of programs.

Greater country ownership of a program through increased financial contribution and

meaningful leadership in designing and administering the program enhances CCT

sustainability. The financial sustainability of conditional cash transfer programs depends on

both the political will of governments to start them, and a broad political consensus to

sustain them to fight poor health, poverty and ensure continued development. Including

improved health outcomes and healthcare utilization along with poverty alleviation thus

needs to be recognized by both political leaders and program managers in the development

and continued support for a CCT program.

Going forward, it is important to consider the definition of outcome indicators, which

currently vary across studies and should be standardized. For instance, “adequate prenatal

care” means different things in different settings, and it was always defined – with the

exception of Mexico and Guatemala – as number of prenatal visits, when existing literature

suggests that the number of visits is not a good predictor of better maternal health

outcomes. Similarly, post-partum visits were defined as a 10-day post-partum checkup in

Honduras; it is not clear whether a 10-day post-partum checkup would be a beneficial

intervention.

It is likely that the quality and availability of supply-side efforts (e.g., skilled care, emergency

care facilities) has major impact, and there is some relationship between including a supply-

side component in the CCT and results. Programs in some states of India, Nicaragua and

Honduras included a supply-side component and –with the exception of Honduras for study

design reasons- report large positive and significant results. However, this issue has not been

examined directly and there are no other examples.

13

One last issue to consider is the cost-effectiveness of these programs, particularly

compared to non-financial incentives. CCTs are not directly comparable with non-financing

demand side approaches as the latter are typically small scale, carried out by non-

governmental groups, and project related. Certain non-financing interventions, such as

maternity waiting homes, integration of traditional birth attendants, birth preparedness and

complication readiness, community referrals, transport systems, and cell phone technologies

to increase use of skilled obstetric care, are promising but require further evaluation [35] [36]

[37] [38]. There are exceptions as Nepal, Pakistan and India, now have means of connecting

women with appropriate MNH services through community workers, either paid or

volunteer. These workers are known to provide household visits and/or hold group

meetings with health talks, and in some parts of these countries they are trained to include

the subjects of birth preparedness, complication readiness and essential newborn care, and

may accompany the woman to a facility for birth. Cost data are patchy at best and need to be

reported for not only the interventions themselves, but for a standard outcome for

comparability (such as DALYs) so that different interventions can be compared in terms of

their cost-effectiveness.

6. Conclusions and Recommendations

Conditional cash transfer programs are increasingly common, particularly narrow CCT

programs which condition specific outcomes such as maternal health, prevention of risky

behavior or vaccination. CCT are particularly gaining popularity in sub-Saharan Africa,

where 18 countries are implementing conditional cash transfer programs [39].

Our review of programs found that CCT have increased the uptake of MNH services,

especially skilled attendance at delivery and antenatal monitoring where consistent results are

reported in a variety of settings: these programs have particularly increased service uptake in

middle-income countries with high income inequality. These effects are seen in both broad

and narrow CCT programs, and considering the time frame of these programs the time-to-

effects can be considered rapid. These results come with three major caveats: rigorous cost-

effectiveness data is not available, main impact channels are not evaluated, and effects are

not directly comparable across different contexts, given varying definitions of poverty and

differences on the supply-side.

We are hoping the evidence we present in this review will be complemented with many more

impact evaluations. 3 of the 18 conditional cash transfer programs in sub-Saharan Africa

have maternal-health related requirements (Eritrea, Mozambique and Senegal). Similarly,

there are ongoing evaluations of maternal health CCT programs in Afghanistan, Bolivia and

the Philippines, and these evaluations can help us understand linkages between utilization

and outcomes.3

3 Fernandez, L. and Olfindo, R. 2011. “Overview of the Philippines Conditional Cash Transfer Program:

The Pantawid Pamilyang Pilipino Program. World Bank, Philippine Social Protection Note. http://www-

14

Regarding the design of both the programs and their evaluations, we recommend the

following for the design of both implementation and evaluation of CCT that target MNH:

Improve evaluation and report standardized outcomes across CCT studies.

Calculate cost-effectiveness estimates for both financial and non-financial

incentives for improved maternal and newborn health.

Focus on the effectiveness and quality of services delivered on the supply side,

in addition to the quantity available.

Pay attention to program design and measure potential impact pathways.

Modify the design to enhance MNH effects with respect to conditionalities,

non-financing incentives and infrastructure barriers.

Add supply-side strengthening conditions to CCT programs; implement

targeted supply-side interventions and track supply-side baseline and outcome

levels.

Understand the link between utilization and outcomes.

wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/2011/07/08/000333038_20110708021205/Render

ed/PDF/628790BRI0Phil0me0abstract0as0no010.pdf

Chavez, Franz. 2010. “Bolivia: Cash for Checkups to Slash Maternal Deaths”

http://www.ipsnews.net/2010/03/bolivia-cash-for-checkups-to-slash-maternal-deaths/

Annex 1. Features of CCT programs in countries where maternal health outcomes were measured

Country Program Name Year Started

Targeting and Eligibility

Number of Beneficiaries

Health Conditions Education Conditions

Verification Supply-Side Conditions & Additional Benefits

Type of Evaluation

Reference(s)

El Salvador

Red Solidaria 2005 Geographic / Proxy Means Testing

100,000 households

Compliance with immunization and regular health and nutrition monitoring

Primary school enrollment / 80% school attendance (5-15 yrs.)

Health and education personnel provide information to NGO

Yes; supply-side component to strengthen basic health and nutrition services in the targeted areas

Regression discontinuity design, differences in differences

de Brauw and Peterman (2011)

Guatemala Mi Familia Progresa

2008 Geographic / Proxy Means Testing

250,000 households

Regular health visits for children (0-16 yr.) and pregnant women

90% of school attendance

Not fully implemented

No Differences in differences

Gutierrez et al (2011)

Honduras Programa de Asignación Familiar

1998 Geographic / Proxy Means Testing

240,000 households

Compliance with required frequency of health center visits; children attend growth monitoring; pregnant women receive at least 4 ANC visits

School enrollment / 85% school attendance

None Yes; promote access to an integrated package of services, including nutrition, healthcare and basic services. Improve quality of facilities due to service-level package

Cluster randomized trial, with a pre-test and post-test cross-section design

Morris, Flores, Olinto and Medina (2004)

India Janani Suraksha Yojana

2005 Poverty-line estimates

9,500,000 women

Delivery in health facility, antenatal checkups

None Community-level health workers

Yes; payments to ASHAs who identify pregnant women and help them get to a facility

Matching, with-versus-without comparison, differences in differences

Lim et al (2010)

Mexico Oportunidades (formerly PROGRESA)

1997 Geographic / Proxy Means Testing

5,000,000 households

Children <2 years fully immunized and undergo growth monitoring. Prenatal visits, breastfeeding, physical checkups

80% school attendance (monthly), and 93% (annually) / Completion of middle school / Completion of grade 12 before age 22

Program state coordination agency

No Regression discontinuity design, differences in differences

Urquieta et al (2009); Stecklov et al (2007); Sosa-Rubi et al (2011); Barber and Gertler (2009) ; Feldman et al (2009) ; Lamadrid-Figueroa et al (2010)

Nepal Safe Delivery Incentive Program (SDIP)

2005 All women 100,000 women

Deliver in a public health facility and had no more than two living children or an obstetric complication. Skilled attendance at birth.

None Deliver in health facility

Yes; provider incentives ($5 for each delivery attended)

Propensity score matching

Powell-Jackson et al (2009); Powell-Jackson et al (2011)

Nicaragua Red de Protección Social

2000 Geographic 3,000 households

Bimonthly health education workshops / Monthly health care visits (aged 0–2) or bimonthly (aged 3–5) / Adequate weight gain and up-to-date vaccinations (aged 0–5)

School enrollment in grades 1-4 (7-13 yrs.) / 85% school attendance (every 2 months) / Grade promotion at end of every year

Forms (confirmed by service providers and put into information system)

Yes; health education workshops every 2 months, child growth and monitoring, provision of antiparasite medicine, vaccinations, teacher transfer

Differences in differences

Stecklov et al (2007)

Uruguay Plan de Atención Nacional a la Emergencia Social

2005 Poverty-line estimates

102,000 individuals

Regular ANC health visits for pregnant women and children

N/A Visits; although not rigorously enforced

No Regression discontinuity design, differences in differences

Amarante et al (2011)

16

Annex 2. Forest plot tables and graphs for interventions

Outcome Country Effect Size Lower CI, 95% Upper CI, 95% % Weight

Adequate prenatal monitoring El Salvador -0.065 -0.206 0.076 7.96

Honduras 0.187 0.069 0.305 10.29

India 0.107 0.091 0.123 29.86

Nepal -0.046 -0.166 0.074 10.07

Mexico 0.081 0.03 0.132 22.47

Guatemala 0.11 -0.021 0.241 8.84

Uruguay 0.144 0.028 0.26 10.52

Overall effect (dl) 0.084 0.038 0.131 100

Birth attended by skilled personnel El Salvador 0.123 -0.014 0.26 15.66

India 0.366 0.354 0.378 17.08

Nepal 0.052 0.021 0.083 17.01

Mexico 0.114 0.02 0.208 16.39

Guatemala 0.04 -0.021 0.101 16.79

Uruguay -0.002 -0.02 0.016 17.07

Overall effect (dl) 0.116 -0.072 0.303 100

Tetanus toxoid for mother Honduras 0.042 -0.098 0.182 89.89

Mexico 0.368 -0.22 0.956 10.11

Overall effect (dl) 0.075 -0.118 0.268 100

Mother gave birth in health facility Nepal 0.04 0.011 0.069 34.08

India 0.435 0.424 0.446 34.16

El Salvador 0.153 0.004 0.302 31.76

Overall effect (dl) 0.211 -0.105 0.527 100

Post-partum checkups/visits after birth Honduras -0.056 -0.157 0.045 79.02

El Salvador -0.059 -0.255 0.137 20.98

Overall effect (dl) -0.057 -0.146 0.033 100

Caesarean section Nepal 0.012 0.001 0.023 80.53

Mexico 0.051 -0.01 0.112 19.47

Overall effect (dl) 0.02 -0.011 0.05 100

Fertility Honduras 0.039 0.012 0.066 8.48

Nicaragua -0.011 -0.142 0.12 0.44

Mexico -0.003 -0.009 0.003 42.93

Uruguay 0.001 -0.003 0.005 48.16

Overall effect (dl) 0.002 -0.006 0.011 100

Low birthweight Mexico -0.046 -0.234 0.142 0.27

Uruguay -0.015 -0.025 -0.005 99.73

Overall effect (dl) -0.015 -0.025 -0.005 100

17

Adequate prenatal monitoring

Skilled attendance at birth

18

Tetanus toxoid vaccination for mother

Gave birth in hospital

19

Post-partum checkups

Caesarean section

20

Fertility

Low birthweight

Annex 3. Overall table of reviewed effects

Study Dependent Variable ES 95% CI

de Brauw and Peterman (2011) (El Salvador)

Adequate prenatal monitoring (5 or more visits) -0.065 (-0.059, 0.065)

Birth attended by skilled personnel 0.123 (0.129, -0.006)

Gave birth in hospital 0.153 (0.147, 0.159)

Mother went for postnatal checkup -0.059 (-0.050, -0.068)

Morris, Flores, Olinto and Medina (2004) (Honduras)

Adequate prenatal monitoring (5 or more visits) 0.187 (0.074, 0.30)

Adequate prenatal monitoring (5 or more visits) 0.184 (0.069, 0.299)

Adequate prenatal monitoring (5 or more visits) 0.132 (-0.016, 0.28)

10-day post-partum check-up -0.056 (-0.157, 0.045)

10-day post-partum check-up -0.057 (-0.16, 0.045)

10-day post-partum check-up 0.012 (-0.118, 0.143)

Child taken to health center at least once in last 30 days 0.202 (0.109, 0.296)

Child taken to health center at least once in last 30 days 0.149 (0.056, 0.243)

Child taken to health center at least once in last 30 days -0.018 (-0.134, 0.098)

Tetanus toxoid for mother 0.042 (-0.097, 0.182)

Tetanus toxoid for mother 0.081 (-0.061, 0.222)

Tetanus toxoid for mother 0.064 (-0.116, 0.244)

Weighed in last 30 days for mother 0.211 (0.111, 0.311)

Weighed in last 30 days for mother 0.176 (0.075, 0.276)

Weighed in last 30 days for mother 0.08 (-0.044, 0.204)

Barber and Gertler (2009) (Mexico)

Tetanus toxoid for mother 0.368 N/A

Average physical examination visits 0.059 N/A

Iron supplements 0.053 N/A

Average increase in prevention and case management 0.043 N/A

Prenatal procedures received 0.122 N/A

Lim et al (2010) (India) Adequate prenatal monitoring (3 or more visits) 0.109 (0.046, 0.172)

Birth attended by skilled personnel 0.393 (0.337, 0.45)

In-facility births 0.492 (0.432, 0.551)

Perinatal deaths (per 1000 pregnancies) -14.2 (-31, 2.7)

Neonatal deaths (per 1000 live births) -6.2 (-20.4, 8.1)

Maternal deaths (per 100000 live births) -100.5 (-582.2, 381.2)

Ozer et al (2011) (Mexico) Full depression scale (0-60) -1.71 (-2.46, -0.96)

Powell-Jackson et al (2011) (Nepal)

Delivery at facility 0.04 0.05,0.31

Birth attended by skilled personnel 0.052 0.06, 0.28

Powell-Jackson et al (2009) (Nepal)

Birth attended by skilled personnel 0.023 (-0.082, 0.129)

Number of antenatal care visits 0.031 (0.008, 0.054)

Neonatal mortality -0.0004 (0,0)

Delivery at home -0.042 (-0.329, 0.245)

Delivery at government facility 0.026 (0.168, -0.116)

Delivery at private facility 0.002 (0.007, -0.003)

Delivery with a health worker 0.044 (0.342, -0.254)

Delivery by caesarean section -0.001 (0, -0.002)

22

Study Dependent Variable ES 95% CI

Sosa-Rubi et al (2011) (Mexico) Antenatal visits 0.021

Baird et al (2011) (Malawi) Teenage pregnancy 0.029 0.027

Teenage pregnancy -0.067 0.024

Stecklov et al (2007) (Latin America)

Fertility; controlled for education, age, household, wealth (Honduras) 0.039 0.002

Fertility; controlled for education, age, household, wealth (Nicaragua) 0.009 0.565

Fertility; controlled for education, age, household, wealth (Mexico) -0.003 0.852

Urquieta et al (2009) (Mexico) Skilled attendance at delivery 0.028 0.027

Alam et al (2010) (Pakistan) Probability of marriage 0.0082 0.008

Age at marriage 1.46 0.621

Probability of giving birth -0.0808 0.172

Number of children -0.329 0.181

Oportunidades, Official Evaluation

Maternal anemia for women of childbearing age, urban 0.003

Maternal anemia for women of childbearing age, rural -0.014

IDB; Gutierrez et al (2011) (Guatemala)

Folic acid supplement 0.07 *

Iron supplement 0.1 **

Number of prenatal visits at health centers 0.11 **

Skilled attendance at delivery 0.01

Baird et al (2009) (Malawi) Risky sexual activity -0.159

Sexual activity (number of partners) -0.036

Teenage pregnancy -0.051 **

Barber (2009) (Mexico) Cesarean section rate 0.0508 **

23

Annex 4. Definition of terms

Low birth weight: The papers cited here use the World Health Organization’s definition of a newborn weighing less than

2,500 grams (5.5 pounds).

Adequate antenatal monitoring: Adequate antenatal monitoring is defined as 5 or more visits to a health facility for

antenatal monitoring, except for the Indian study, which defines it as 3 or more visits to a health facility for antenatal

monitoring.

Births attended by skilled personnel: Skilled attendance at birth is defined as attendance by a doctor, an

obstetrician/gynecologist, a nurse or a midwife.

Births in health facility: Studies define “facility” differently; the El Salvador study includes births in a public or private

facility and excludes births at health centers or mobile health clinics, the Indian study includes any kind of health facility.

Post-partum visits: A visit to a health facility within 10-14 days of giving birth.

24

References

1. Ronsmans, C. and W.J. Graham, Maternal mortality: who, when, where, and why. The Lancet, 2006. 368(9542):

p. 1189-1200.

2. Darmstadt, G.L., et al., Evidence-based, cost-effective interventions: how many newborn babies can we save? The

Lancet, 2005. 365(9463): p. 977-988.

3. UNICEF, The state of the world's children, 2008: New York.

4. Barros, A.J.D., et al., Equity in maternal, newborn, and child health interventions in Countdown to 2015: a

retrospective review of survey data from 54 countries. The Lancet, 2012. 379(9822): p. 1225-1233.

5. World Bank, World Development Indicators, 2012.

6. Ensor, T. and S. Cooper, Overcoming barriers to health service access: influencing the demand side. Health Policy

Plan, 2004. 19(2): p. 69-79.

7. Lassi Zohra, S., A. Haider Batool, and A. Bhutta Zulfiqar Community-based intervention packages for reducing

maternal and neonatal morbidity and mortality and improving neonatal outcomes. Cochrane Database of

Systematic Reviews, 2010. DOI: 10.1002/14651858.CD007754.pub2.

8. Gaarder, M.G., Amanda; Todd, Jessica E., Conditional cash transfers and health: unpacking the causal chain.

Journal of Development Effectiveness, 2010. 2(1).

9. Witter, S.K., Sunil; Nath, Hom; Tiwari, Suresh, The national free delivery policy in Nepal: early evidence of its

effects on health facilities. Health Policy and Planning, 2011. 26(suppl 2).

10. Regalia, F.C., Leslie, Nicaragua: Combining Demand- and Supply-Side Incentives, in Performance Incentives for

Global Health, R. Eichler, Levine, Ruth, Editor 2009, Center for Global Development: Washington, DC.

11. USAID. Paying for performance: The Janani Suraksha Yojana Program in India. 2012 [cited 2012; Available from:

http://www.healthsystems2020.org/content/resource/detail/2609/.

12. de Brauw, A. and A. Peterman, Can conditional cash transfers improve maternal health and birth outcomes?

Evidence from El Salvador’s Comunidades Solidarias Rurales, in IFPRI Discussion Paper2011, International Food

Policy Research Institute. p. 36.

13. Morris, S.S., et al., Monetary incentives in primary health care and effects on use and coverage of preventive

health care interventions in rural Honduras: cluster randomised trial. The Lancet, 2004. 364(9450): p. 2030-2037.

14. Lim, S.S., et al., India's Janani Suraksha Yojana, a conditional cash transfer programme to increase births in

health facilities: an impact evaluation. Lancet, 2010. 375(9730): p. 2009-23.

25

15. Powell-Jackson, T. and K. Hanson, Financial incentives for maternal health: Impact of a national programme in

Nepal. Journal of Health Economics, 2010.

16. Barber, S.L. and P.J. Gertler, Empowering women to obtain high quality care: evidence from an evaluation of

Mexico's conditional cash transfer programme. Health Policy Plan, 2009. 24(1): p. 18-25.

17. Gutiérrez, J.P., Evaluación Externa de Impacto del Programa de Transferencias Monetarias Condicionadas, 2011,

Banco Interamericano de Desarrollo. p. 309.

18. Amarante, V., et al., Do Cash Transfers Improve Birth Outcomes? Evidence from Matched Vital Statistics, Social

Security and Program Data, in NBER Working Paper 2011, The National Bureau of Economic Research. p. 44.

19. Urquieta, J., et al., Impact of oportunidades on skilled attendance at delivery in rural areas. Economic

Development and Cultural Change, 2009. 57(3): p. 539-558.

20. Feldman, B.S., et al., Contraceptive Use, Birth Spacing, and Autonomy: An Analysis of the Oportunidades

Program in Rural Mexico. Studies in Family Planning, 2009. 40(1): p. 51-62.

21. Lamadrid-Figueroa, H., et al., Heterogeneous impact of the social programme Oportunidades on use of

contraceptive methods by young adult women living in rural areas. Journal of Development Effectiveness, 2010.

2(1): p. 74-86.

22. Stecklov, G., et al., Unintended effects of poverty programmes on childbearing in less developed countries:

Experimental evidence from Latin America. Population Studies, 2007. 61(2): p. 125-140.

23. Hernandez, B., Ramirez, D., Moreno, H., & Laird, N. , Evaluacion del impacto de Oportunidades en la mortalidad

materna e infantile. In Resultados de la Evaluacion Externa del Programa de Desarrollo Humano Oportunidades.

[Results of the External Evaluation of Programa de Desarrollo Humano Oportunidades.], 2004: Cuernavaca,

Mexico.

24. Mazumdar, S., A. Mills, and T. Powell-Jackson, Financial Incentives in Health: New Evidence from India's Janani

Suraksha Yojana, 2011.

25. Das, A., D. Rao, and A. Hagopian, India's Janani Suraksha Yojana: further review needed. The Lancet, 2011.

377(9762): p. 295-296.

26. Devadasan, N., et al., A conditional cash assistance programme for promoting institutional deliveries among the

poor in India: process evaluation results, in Reducing financial barriers to obstetric care in low-income countries,

R. F., W. S., and D.B. V., Editors. 2008, ITG Press: Belgium.

27. Triyana, M., Do Health Care Providers Respond to Demand-Side Incentives? Evidence from Indonesia, 2012,

University of Chicago.

26

28. Barber, S.L. and P.J. Gertler, The impact of Mexico's conditional cash transfer programme, Oportunidades, on

birthweight. Trop Med Int Health, 2008. 13(11): p. 1405-14.

29. Carpio, A.A.J.E.B.X.V.D., Does Cash for School Influence Young Women’s Behavior in the Longer Term? Evidence

from Pakistan, in World Bank Policy Research Working Papers2011.

30. Baird, S., et al., The Short-Term Impacts of a Schooling Conditional Cash Transfer Program on the Sexual Behavior

of Young Women, in Impact Evaluation Series2009, The World Bank: Washington, D.C. p. 33.

31. Baird, S., et al., Effect of a cash transfer programme for schooling on prevalence of HIV and herpes simplex type 2

in Malawi: a cluster randomised trialThe Lancet, 2012. 379(9823): p. 1320-1329.

32. Slater, R., J. Farrington, and R. Holmes, A Conceptual Framework for Understanding the Role of Cash Transfers in

Social Protection ODI, Editor 2008: London.

33. Adato, M. and T. Roopnaraine, Women’s status, gender relations and conditional cash transfers. , in Conditional

Cash Transfers in Latin America, M. Adato and J. Hoddinott, Editors. 2010, The Johns Hopkins University Press

Washington DC.

34. Sosa-Rubí, S.G., et al., Learning effect of a conditional cash transfer programme on poor rural women's selection

of delivery care in Mexico. Health Policy and Planning, 2011. 26(6): p. 496-507.

35. Byrne, A. and A. Morgan, How the integration of traditional birth attendants with formal health systems can

increase skilled birth attendance. International journal of gynaecology and obstetrics: the official organ of the

International Federation of Gynaecology and Obstetrics, 2011. 115(2): p. 127-134.

36. Varkey, L., et al., Involving Men in Maternity Care in India, 2004, Population Council. p. 72.

37. Hussein, J., et al., What kinds of policy and programme interventions contribute to reductions in maternal

mortality? The effectiveness of primary level referral systems for emergency maternity care in developing

countries, 2011, London: EPPI-Centre, Social Science Research Unit, Institute of Education, University of London:

London. p. 54.

38. Holmes, W. and E. Kennedy, Reaching emergency obstetric care: overcoming the ‘second delay’, 2010, Burnet

Institute on behalf of Compass, the Women’s and Children’s Health Knowledge Hub: Melbourne, Australia. p.

105.

39. Garcia, M. and C.M.T. Moore, The Cash Dividend : The Rise of Cash Transfer Programs in Sub-Saharan Africa, in

Directions in Development2012, The World Bank: Washington D.C. p. 440.